webapp.py 12 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364
  1. #!/usr/bin/env python
  2. '''
  3. searx is free software: you can redistribute it and/or modify
  4. it under the terms of the GNU Affero General Public License as published by
  5. the Free Software Foundation, either version 3 of the License, or
  6. (at your option) any later version.
  7. searx is distributed in the hope that it will be useful,
  8. but WITHOUT ANY WARRANTY; without even the implied warranty of
  9. MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  10. GNU Affero General Public License for more details.
  11. You should have received a copy of the GNU Affero General Public License
  12. along with searx. If not, see < http://www.gnu.org/licenses/ >.
  13. (C) 2013- by Adam Tauber, <asciimoo@gmail.com>
  14. '''
  15. import json
  16. import cStringIO
  17. import os
  18. from flask import Flask, request, render_template
  19. from flask import url_for, Response, make_response, redirect
  20. from flask import send_from_directory
  21. from searx import settings, searx_dir
  22. from searx.engines import search
  23. from searx.engines import categories
  24. from searx.engines import engines
  25. from searx.engines import get_engines_stats
  26. from searx.engines import engine_shortcuts
  27. from searx.utils import UnicodeWriter
  28. from searx.utils import highlight_content, html_to_text
  29. from searx.languages import language_codes
  30. from flask.ext.babel import Babel
  31. app = Flask(
  32. __name__,
  33. static_folder=os.path.join(searx_dir, 'static'),
  34. template_folder=os.path.join(searx_dir, 'templates')
  35. )
  36. app.secret_key = settings['server']['secret_key']
  37. babel = Babel(app)
  38. #TODO configurable via settings.yml
  39. favicons = ['wikipedia', 'youtube', 'vimeo', 'soundcloud',
  40. 'twitter', 'stackoverflow', 'github']
  41. @babel.localeselector
  42. def get_locale():
  43. locale = request.accept_languages.best_match(settings['locales'].keys())
  44. if request.cookies.get('locale', '') in settings['locales']:
  45. locale = request.cookies.get('locale', '')
  46. if 'locale' in request.args\
  47. and request.args['locale'] in settings['locales']:
  48. locale = request.args['locale']
  49. if 'locale' in request.form\
  50. and request.form['locale'] in settings['locales']:
  51. locale = request.form['locale']
  52. return locale
  53. def get_base_url():
  54. if settings['server']['base_url']:
  55. hostname = settings['server']['base_url']
  56. else:
  57. scheme = 'http'
  58. if request.is_secure:
  59. scheme = 'https'
  60. hostname = url_for('index', _external=True, _scheme=scheme)
  61. return hostname
  62. def render(template_name, **kwargs):
  63. global categories
  64. kwargs['categories'] = ['general']
  65. kwargs['categories'].extend(x for x in
  66. sorted(categories.keys()) if x != 'general')
  67. if not 'selected_categories' in kwargs:
  68. kwargs['selected_categories'] = []
  69. cookie_categories = request.cookies.get('categories', '').split(',')
  70. for ccateg in cookie_categories:
  71. if ccateg in categories:
  72. kwargs['selected_categories'].append(ccateg)
  73. if not len(kwargs['selected_categories']):
  74. kwargs['selected_categories'] = ['general']
  75. return render_template(template_name, **kwargs)
  76. def parse_query(query):
  77. query_engines = []
  78. query_parts = query.split()
  79. if query_parts[0].startswith('!'):
  80. prefix = query_parts[0][1:].replace('_', ' ')
  81. if prefix in engine_shortcuts:
  82. query_engines.append({'category': 'none',
  83. 'name': engine_shortcuts[prefix]})
  84. elif prefix in engines:
  85. query_engines.append({'category': 'none',
  86. 'name': prefix})
  87. elif prefix in categories:
  88. query_engines.extend({'category': prefix,
  89. 'name': engine.name}
  90. for engine in categories[prefix])
  91. if len(query_engines):
  92. query = query.replace(query_parts[0], '', 1).strip()
  93. return query, query_engines
  94. @app.route('/', methods=['GET', 'POST'])
  95. def index():
  96. """Render index page.
  97. Supported outputs: html, json, csv, rss.
  98. """
  99. paging = False
  100. lang = 'all'
  101. if request.cookies.get('language')\
  102. and request.cookies['language'] in (x[0] for x in language_codes):
  103. lang = request.cookies['language']
  104. if request.method == 'POST':
  105. request_data = request.form
  106. else:
  107. request_data = request.args
  108. if not request_data.get('q'):
  109. return render('index.html')
  110. pageno_param = request_data.get('pageno', '1')
  111. if not pageno_param.isdigit() or int(pageno_param) < 1:
  112. return render('index.html')
  113. pageno = int(pageno_param)
  114. selected_categories = []
  115. query, selected_engines = parse_query(request_data['q'].encode('utf-8'))
  116. if len(selected_engines):
  117. selected_categories = list(set(engine['category']
  118. for engine in selected_engines))
  119. else:
  120. for pd_name, pd in request_data.items():
  121. if pd_name.startswith('category_'):
  122. category = pd_name[9:]
  123. if not category in categories:
  124. continue
  125. selected_categories.append(category)
  126. if not len(selected_categories):
  127. cookie_categories = request.cookies.get('categories', '')
  128. cookie_categories = cookie_categories.split(',')
  129. for ccateg in cookie_categories:
  130. if ccateg in categories:
  131. selected_categories.append(ccateg)
  132. if not len(selected_categories):
  133. selected_categories = ['general']
  134. for categ in selected_categories:
  135. selected_engines.extend({'category': categ,
  136. 'name': x.name}
  137. for x in categories[categ])
  138. results, suggestions = search(query,
  139. request,
  140. selected_engines,
  141. pageno,
  142. lang)
  143. for result in results:
  144. if not paging and engines[result['engine']].paging:
  145. paging = True
  146. if request_data.get('format', 'html') == 'html':
  147. if 'content' in result:
  148. result['content'] = highlight_content(result['content'], query)
  149. result['title'] = highlight_content(result['title'], query)
  150. else:
  151. if 'content' in result:
  152. result['content'] = html_to_text(result['content']).strip()
  153. result['title'] = html_to_text(result['title']).strip()
  154. if len(result['url']) > 74:
  155. url_parts = result['url'][:35], result['url'][-35:]
  156. result['pretty_url'] = '{0}[...]{1}'.format(*url_parts)
  157. else:
  158. result['pretty_url'] = result['url']
  159. for engine in result['engines']:
  160. if engine in favicons:
  161. result['favicon'] = engine
  162. if request_data.get('format') == 'json':
  163. return Response(json.dumps({'query': query, 'results': results}),
  164. mimetype='application/json')
  165. elif request_data.get('format') == 'csv':
  166. csv = UnicodeWriter(cStringIO.StringIO())
  167. keys = ('title', 'url', 'content', 'host', 'engine', 'score')
  168. if len(results):
  169. csv.writerow(keys)
  170. for row in results:
  171. row['host'] = row['parsed_url'].netloc
  172. csv.writerow([row.get(key, '') for key in keys])
  173. csv.stream.seek(0)
  174. response = Response(csv.stream.read(), mimetype='application/csv')
  175. content_disp = 'attachment;Filename=searx_-_{0}.csv'.format(query)
  176. response.headers.add('Content-Disposition', content_disp)
  177. return response
  178. elif request_data.get('format') == 'rss':
  179. response_rss = render(
  180. 'opensearch_response_rss.xml',
  181. results=results,
  182. q=request_data['q'],
  183. number_of_results=len(results),
  184. base_url=get_base_url()
  185. )
  186. return Response(response_rss, mimetype='text/xml')
  187. return render(
  188. 'results.html',
  189. results=results,
  190. q=request_data['q'],
  191. selected_categories=selected_categories,
  192. paging=paging,
  193. pageno=pageno,
  194. suggestions=suggestions
  195. )
  196. @app.route('/about', methods=['GET'])
  197. def about():
  198. """Render about page"""
  199. return render('about.html')
  200. @app.route('/engines', methods=['GET'])
  201. def list_engines():
  202. """Render engines page.
  203. List of all supported engines.
  204. """
  205. global categories
  206. return render('engines.html', categs=categories.items())
  207. @app.route('/preferences', methods=['GET', 'POST'])
  208. def preferences():
  209. """Render preferences page.
  210. Settings that are going to be saved as cookies."""
  211. lang = None
  212. if request.cookies.get('language')\
  213. and request.cookies['language'] in (x[0] for x in language_codes):
  214. lang = request.cookies['language']
  215. if request.method == 'POST':
  216. selected_categories = []
  217. locale = None
  218. for pd_name, pd in request.form.items():
  219. if pd_name.startswith('category_'):
  220. category = pd_name[9:]
  221. if not category in categories:
  222. continue
  223. selected_categories.append(category)
  224. elif pd_name == 'locale' and pd in settings['locales']:
  225. locale = pd
  226. elif pd_name == 'language' and (pd == 'all' or
  227. pd in (x[0] for
  228. x in language_codes)):
  229. lang = pd
  230. resp = make_response(redirect('/'))
  231. if locale:
  232. # cookie max age: 4 weeks
  233. resp.set_cookie(
  234. 'locale', locale,
  235. max_age=60 * 60 * 24 * 7 * 4
  236. )
  237. if lang:
  238. # cookie max age: 4 weeks
  239. resp.set_cookie(
  240. 'language', lang,
  241. max_age=60 * 60 * 24 * 7 * 4
  242. )
  243. if selected_categories:
  244. # cookie max age: 4 weeks
  245. resp.set_cookie(
  246. 'categories', ','.join(selected_categories),
  247. max_age=60 * 60 * 24 * 7 * 4
  248. )
  249. return resp
  250. return render('preferences.html',
  251. locales=settings['locales'],
  252. current_locale=get_locale(),
  253. current_language=lang or 'all',
  254. language_codes=language_codes)
  255. @app.route('/stats', methods=['GET'])
  256. def stats():
  257. """Render engine statistics page."""
  258. global categories
  259. stats = get_engines_stats()
  260. return render('stats.html', stats=stats)
  261. @app.route('/robots.txt', methods=['GET'])
  262. def robots():
  263. return Response("""User-agent: *
  264. Allow: /
  265. Allow: /about
  266. Disallow: /stats
  267. Disallow: /engines
  268. """, mimetype='text/plain')
  269. @app.route('/opensearch.xml', methods=['GET'])
  270. def opensearch():
  271. method = 'post'
  272. # chrome/chromium only supports HTTP GET....
  273. if request.headers.get('User-Agent', '').lower().find('webkit') >= 0:
  274. method = 'get'
  275. ret = render('opensearch.xml', method=method, host=get_base_url())
  276. resp = Response(response=ret,
  277. status=200,
  278. mimetype="application/xml")
  279. return resp
  280. @app.route('/favicon.ico')
  281. def favicon():
  282. return send_from_directory(os.path.join(app.root_path, 'static/img'),
  283. 'favicon.png',
  284. mimetype='image/vnd.microsoft.icon')
  285. def run():
  286. from gevent import monkey
  287. monkey.patch_all()
  288. app.run(
  289. debug=settings['server']['debug'],
  290. use_debugger=settings['server']['debug'],
  291. port=settings['server']['port']
  292. )
  293. if __name__ == "__main__":
  294. run()